Karine Megerdoomian: Unaccusatives and special meaning (reply to Martha McGinnis)

[Martha McGinnis]

Hello,

>In their book "Unaccusativity," Levin & Rappaport-Hovav argue that
>verbs undergoing the causative alternation are underlyingly
>transitive.  ...

I'm coming into this discussion a bit late I guess but I have been working
on causative alternation verbs in Armenian recently and I believe that
there is enough evidence to show that (at least in this language) some
verbs are underlyingly transitive. These verb categories don't exactly
correspond to Levin and Rappaport Hovav's verbs though. So, for instance, a
verb such as "break", "cook", "sink" would belong to the underlyingly
transitive category but verbs such as "dry", "widen", "grow" do not.

The difference between the two verbs comes from various morphological and
syntactic phenomena. Morphologically, to transitivize the verb "dry" an
overt causative morpheme "ts(n/r)" needs to be added to the verb (1). On
the other hand, the transitive form of "break" does not have any overt
morphology but its intransitive form has a passive/reflexive morpheme (2).

(1) mirk-e  chor-a-tsr-etsi
     fruit-acc dry-Th-Caus-past.1sg      [Th=thematic vowel]
     'I dried the fruit'

(2) a. bajak-e k'ot'r-etsi
        glass-acc break-past.1sg
       'I broke the glass'
     b. bajak'-e k'ot'r-v-ets
        glass-nom break-Pass/Refl-past.1sg
       'The glass broke'

But more importantly, the syntactic/semantic behavior of the two verb
categories are also different. So for instance,  in forming an adjectival
participle, when the morpheme "atz" (Resultative)is added on the
intransitive form of "dry", there is no causer present but if it is added
on the causative version, then a causer is implied in the event (3).
The "break" verb in (4), in its morphologically unmarked form clearly
corresponds to the causative reading in (3b)

(3) a. chor-ats-atz mirk-e
        dry-Aorist-Res fruit-nom
        'the dried fruit' (only result is known; no causer)
     b. chor-a-tsr-atz mirk-e
        dry-Th-Caus-Res fruit-nom
         'the dried fruit' (someone caused the fruit to become dry)

(4) k'ot'r-atz bajak'-e
     break-Res glass-nom
     'the broken glass' (someone caused the glass to break)

Based on these examples and other syntactic tests, I think that the "break"
verbs are underlyingly transitive/causative. But I also think that this may
vary from one language to another. For instance, in Persian, verbs such
as "sink" or "dry" are formed the same way:

(5) a. xoshk/qarq  shodan
        dry/sink become
        'dry/sink (intr)'
     b. xoshk/qarq  kardan
        dry/sink make
        'dry/sink (tr.)'

I think the idea that some verbs are underlyingly transitive is a valid
problem to investigate since, as Martha McGinnis points out, it raises a
problem for the DM approach. I'd love to hear your comments and suggestions.

-Karine